mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] squashfs: fix NULL pointer dereference in bio_alloc_clone failure path
@ 2025-05-15 12:01 caius.zone
  2025-05-15 12:41 ` Matthew Wilcox
  0 siblings, 1 reply; 2+ messages in thread
From: caius.zone @ 2025-05-15 12:01 UTC (permalink / raw)
  To: phillip; +Cc: linux-fsdevel, linux-kernel, Caius Zone

From: Caius Zone <zone@ai-sast.com>

bio_alloc_clone() may return NULL under memory pressure. The current code
does not check its return value, which may lead to a NULL pointer dereference
in bio_chain() or other bio operations.

Add a NULL check and return -ENOMEM if allocation fails.

Signed-off-by: Caius Zone <zone@ai-sast.com>
---
 fs/squashfs/block.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/fs/squashfs/block.c b/fs/squashfs/block.c
index 2dc730800f44..b00a71f8933c 100644
--- a/fs/squashfs/block.c
+++ b/fs/squashfs/block.c
@@ -115,6 +115,9 @@ static int squashfs_bio_read_cached(struct bio *fullbio,
 			struct bio *new = bio_alloc_clone(bdev, fullbio,
 							  GFP_NOIO, &fs_bio_set);
 
+			if (!new)
+				return -ENOMEM;
+
 			if (bio) {
 				bio_trim(bio, start_idx * PAGE_SECTORS,
 					 (end_idx - start_idx) * PAGE_SECTORS);
-- 
2.25.1


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH] squashfs: fix NULL pointer dereference in bio_alloc_clone failure path
  2025-05-15 12:01 [PATCH] squashfs: fix NULL pointer dereference in bio_alloc_clone failure path caius.zone
@ 2025-05-15 12:41 ` Matthew Wilcox
  0 siblings, 0 replies; 2+ messages in thread
From: Matthew Wilcox @ 2025-05-15 12:41 UTC (permalink / raw)
  To: caius.zone; +Cc: phillip, linux-fsdevel, linux-kernel, Caius Zone

On Thu, May 15, 2025 at 08:01:54PM +0800, caius.zone@icloud.com wrote:
> diff --git a/fs/squashfs/block.c b/fs/squashfs/block.c
> index 2dc730800f44..b00a71f8933c 100644
> --- a/fs/squashfs/block.c
> +++ b/fs/squashfs/block.c
> @@ -115,6 +115,9 @@ static int squashfs_bio_read_cached(struct bio *fullbio,
>  			struct bio *new = bio_alloc_clone(bdev, fullbio,
>  							  GFP_NOIO, &fs_bio_set);
>  
> +			if (!new)
> +				return -ENOMEM;
> +
>  			if (bio) {
>  				bio_trim(bio, start_idx * PAGE_SECTORS,
>  					 (end_idx - start_idx) * PAGE_SECTORS);

This is definitely wrong.  The check for 'new' being NULL should be
after the submit_bio() otherwise we'll leak the previous bio.

But then we need to not call bio_chain() ... so I think this ends up
looking like:

@@ -115,12 +115,15 @@ static int squashfs_bio_read_cached(struct bio *fullbio,
                        struct bio *new = bio_alloc_clone(bdev, fullbio,
                                                          GFP_NOIO, &fs_bio_set);

+                       if (new)
+                               bio_chain(bio, new);
                        if (bio) {
                                bio_trim(bio, start_idx * PAGE_SECTORS,
                                         (end_idx - start_idx) * PAGE_SECTORS);
-                               bio_chain(bio, new);
                                submit_bio(bio);
                        }
+                       if (!new)
+                               return -ENOMEM;

                        bio = new;
                        start_idx = idx;

... but that might not be right either.  We might need to break out of
the loop, rather than returning and do the submit_bio_wait() followed by
the bio_put().

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2025-05-15 12:41 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-05-15 12:01 [PATCH] squashfs: fix NULL pointer dereference in bio_alloc_clone failure path caius.zone
2025-05-15 12:41 ` Matthew Wilcox

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®